[Breaking] Remove DualizableModel

[blegat]

Testing whether it would work to just remove it

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 94.20%. Comparing base (c347f54) to head (2c2135a).
⚠️ Report is 6 commits behind head on master.

Additional details and impacted files

@@           Coverage Diff           @@
##           master     #210   +/-   ##
=======================================
  Coverage   94.20%   94.20%           
=======================================
  Files          14       14           
  Lines         742      742           
=======================================
  Hits          699      699           
  Misses         43       43           

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[joaquimg]

the idea is that it would block certain cones, isn't it?
what is the alternative? let it error?

[blegat]

the idea is that it would block certain cones, isn't it?
what is the alternative? let it error?

If there is a MOI.Utilities.UniversalFallback then it will report that anything is supported anyway.
If a cone is not implementing MOI.dual_set, we already have supports_constraint return the right thing:

Dualization.jl/src/MOI_wrapper.jl

Lines 146 to 149 in c347f54

	D = _dual_set_type(S)
	if D === nothing
	return false
	end

Dualization.jl/src/dual_sets.jl

Lines 10 to 16 in c347f54

	function _dual_set_type(::Type{S}) where {S}
	return try
	MOI.dual_set_type(S)
	catch
	return nothing # The fallback of `dual_set_type` throws an error.
	end
	end

The advantage with that approach is that one can extend Dualization with new cones as well.